Communication Dans Un Congrès Année : 2025

GuqinSonGest: A multimodal and historically grounded dataset of guqin playing techniques

Résumé

This paper introduces GuqinSonGest, a multimodal dataset of guqin isolated playing techniques which is derived from a 9th century Chinese treatise, the Tang Chenzhuo Zhifa (唐陈拙指法). Guqin-SonGest is a dataset for comparative analysis: it contains audio and video samples of playing techniques subject to elementary gestural variations, which we denote by subtechniques. We motivate the construction of GuqinSonGest by emphasizing the significant yet ambiguous gesture-timbre relationships in guqin compositions, pedagogy, and historical guqin writings. We accumulate around three hours of audiovisual materials, containing samples from five families of playing techniques with systematically varied subtechniques and a full-length guqin piece. We validate the timbral variability caused by differing gestural subtechniques by computationally testing a few gesture-timbre hypothesis.

Fichier principal
Vignette du fichier
han2025smc.pdf (5.9 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05095944 , version 1 (03-06-2025)

Licence

Identifiants

  • HAL Id : hal-05095944 , version 1

Citer

Han Han, Vincent Lostanlen, Mathieu Lagrange. GuqinSonGest: A multimodal and historically grounded dataset of guqin playing techniques. Sound and Music Computing Conference (SMC), Jul 2025, Graz, Austria. ⟨hal-05095944⟩
166 Consultations
368 Téléchargements

Partager

  • More